Although community-level restrictions on alcohol availability to youth are important local intervention strategies, little is known about how variations in local physical availability of alcohol are related to: (a) alcohol consumption and problems among young people, (b) the use of alternative sources of alcohol, or (c) drinking trajectories over time. Moreover, the processes through which restrictions in alcohol availability affect drinking by young people are unknown. Differences in the physical availability of alcohol may influence drinking and related problems directly or may be mediated through beliefs about the ease of obtaining alcohol (subjective availability), beliefs about the personal consequences of drinking (alcohol expectancies), or perceptions of drinking norms (normative beliefs). Alternatively, physical availability may affect drinking and drinking-related problems among young people because it moderates the relationships between predispositions to drink and alcohol consumption. In particular, lower availability may weaken these relationships. In order to gain better understanding of the role of physical availability of alcohol in adolescent drinking, it proposed to undertake a 3 year longitudinal random digit dial telephone survey study of 1,500 young people (ages 15-17 at Wave 1) sampled from 50 zip codes in the state of California that are known to differ in alcohol outlet density and socio-economic status. The study will proceed from a well-developed conceptual model that includes community-level variables (outlet density, population density, SES, ethnic composition, community disorganization) as well as individual-level variables (drinking, drinking beliefs, personal risk factors, background characteristics). The model further specifies how the effects of community variables are mediated through and possibly moderate the effects of drinking predispositions. Multi-level latent variable structural equations modeling and multi-level latent variable growth modeling techniques will be used to test this conceptual model. The ultimate goal of this research is to (a) provide a better understanding of how physical availability of alcohol in the form of outlet density affects adolescent drinking and drinking problems and thereby (b) provide a better basis for designing and implementing more effective interventions to reduce and prevent adolescent drinking. Ultimately, this information will help policy makers and community advocates make better decisions about local prevention activities and the allocation of scarce prevention resources.